Cancer is the second leading cause of death worldwide (1), and its incidence continues to rise annually (2). It is estimated that by 2040, there will be ~290,000 new cancer diagnoses and ~50,000 cancer-related deaths, with >16% of deaths occurring in low- and middle-income countries (3,4). Cancer treatment imposes substantial physical and psychological suffering on patients and places a marked financial burden on them (5,6). This multifaceted issue expands beyond individual patients, affecting families, careers and the broader society (7). Consequently, numerous studies are actively investigating risk factors associated with cancer (8). A comprehensive understanding of cancer etiology, along with effective interventions, is urgently needed.

Inflammation, particularly chronic inflammation, is a known contributor to tumor development (9-11). This association is supported by multiple mechanisms through which inflammation influences key processes in tumorigenesis, including genomic instability, DNA damage, oncogene activation and impaired tumor suppressor activity (12). Beyond contributing to the early stages of cancer development, chronic inflammation also serves a pivotal role in cancer progression (13,14); ~25% of all cancers are associated with chronic inflammation (15).

Among several peripheral inflammatory markers, high-sensitivity C reactive protein (hs-CRP) has been extensively studied due to its long plasma half-life, minimal diurnal variation and independence from age or sex (16). A previous prospective cohort study reported a notable association between elevated hs-CRP levels and cancer risk (17). Moreover, our previous studies have further supported this association, demonstrating a link between elevated hs-CRP levels and unfavorable survival outcomes across multiple cancer types (18,19). Collectively, these studies indicate a strong association between hs-CRP and cancer; however, a key limitation is that hs-CRP levels were measured only once at baseline.

Nonetheless, cancer development is often associated with the long-term effects of risk factors. As an acute-phase response protein, hs-CRP is susceptible to fluctuations triggered by both endogenous and exogenous factors, particularly in the presence of inflammation, autoimmune disease or cancer (16). Relying solely on baseline hs-CRP levels as a predictor of cancer development has limitations. By contrast, serial measurements of hs-CRP over time, capturing both high and low levels as well as their fluctuations, provide a comprehensive reflection of peripheral inflammation and the associated inflammatory response. Variability, defined as the degree of fluctuation in a biomarker over time, constitutes a critical metric. Commonly used variability indicators include the coefficient of variation (CV) and the variation independent of the mean (VIM). Therefore, the present study aimed to assess changes in hs-CRP over time, expressing variability using the CV, and incorporating VIM as a complementary indicator to explore the association between hs-CRP variability and cancer incidence. The Kailuan cohort has previously been used to examine CRP trajectories and cancer risk (20). The present study uses the same source cohort and the same 2006-2010 CRP measurement period but addresses a different scientific question. It focuses on CRP variability rather than trajectory patterns.

The present study included 42,514 participants from the Kailuan cohort. The detailed process of participant inclusion and exclusion is shown in Fig. 1. Among the included participants, 24% were women and 76% were men, with an average age of 53.24 (11.69) years and a median follow-up duration of 10.98 years. Participants with higher hs-CRP variability tended to be men, younger and have lower BMIs (Table I). Mean hs-CRP was categorized as high (≥3 mg/l) or low (<3 mg/l) based on clinical experience (23), and participants were stratified according to present baseline information. The results revealed that participants with higher hs-CRP were more likely to be women, older and have higher BMIs (Table SI).

Cox regression analysis was then performed to evaluate the association between hs-CRP variability and cancer occurrence. The high hs-CRP group had a ~20% higher risk of cancer than the low hs-CRP group [hazard ratio (HR), 1.20; 95% confidence interval (CI), 1.07-1.33; P<0.001]. RCS was then used to observe the association between continuous CV of hs-CRP and cancer risk, and the results demonstrated that as CV increased, HR decreased, suggesting that higher visit-to-visit variability of hs-CRP was associated with lower cancer risk (Fig. 2A). Subsequently, hs-CRP variability was stratified into three groups based on the tertile intervals of CV of hs-CRP: Low, medium and high. Moderate hs-CRP variability did not show a significant association with cancer risk compared with low variability (adjusted HR, 0.93; 95% CI, 0.84-1.04; P=0.205), but high variability exhibited a notable protective effect against cancer (adjusted HR, 0.84; 95% CI, 0.75-0.94; P=0.003). To further assess the influence of hs-CRP levels on the relationship between variability and cancer risk, an additional Cox regression analysis was performed by cross-classifying hs-CRP levels and variability. Using individuals with low hs-CRP levels and high variability as the reference group, all other groups showed elevated cancer risk compared with the reference (HR >1 for each group), indicating that the reference group had the lowest cancer risk. Compared with that of the reference group, the group with low hs-CRP level and low variability had a ~20% increased risk of cancer (adjusted HR, 1.20; 95% CI, 1.06-1.37; P=0.05), and the group with low hs-CRP levels and moderate variability did not have a significantly different cancer risk (adjusted HR, 1.12; 95% CI, 0.99-1.28; P=0.079). High hs-CRP levels were associated with a higher risk of cancer occurrence regardless of variability, with the group exhibiting the lowest variability showing the highest risk (adjusted HR, 1.42; 95% CI, 1.18-1.70; P<0.001; model C; Table II).

Subsequently, participants were categorized into three distinct long-term hs-CRP change patterns: Stable, increasing and decreasing (Fig. S1). The association between hs-CRP variability and cancer risk in participants with each of these three different patterns was then assessed. The results indicated that only in the stable group, hs-CRP variability was significantly associated with the risk of cancer (high variability in stable-pattern: HR, 0.81; 95% CI, 0.72-0.92; P=0.001; Table III).

Furthermore, the association between hs-CRP variability and cancer at different specific sites was evaluated. The incidence of cancer at several specific sites during the follow-up period is presented in Table SII. Among all participants, a total of 502 incident lung cancer cases occurred during the follow-up period (Table SII). Table IV indicated a significant association between CV of hs-CRP and the incidence of lung cancer (HR, 0.77; 95% CI, 0.63-0.96; P=0.018; Table IV).

To explore whether the association between hs-CRP variability and cancer exists in populations with different characteristics, subgroup analyses were performed. In most subgroups, high variability was associated with an elevated cancer risk. However, the P-values for interaction were not significant (all P>0.05), suggesting no significant heterogeneity across subgroups. Moreover, high variability was more protective against cancer than moderate variability (Fig. 3).

Additionally, sensitivity analyses were performed to ensure the stability of the findings and to rule out other factors. To minimize the possibility of reverse causation, lag analyses were performed by excluding participants who developed cancer within the first 2 years after the exposure assessment period. The results remained consistent with the main findings. The adjusted HR for the highest vs. lowest tertile of hs-CRP-CV was 0.84 (95% CI, 0.75-0.94) in the primary analysis and 0.79 (95% CI, 0.70-0.90) after excluding early cancer cases in the highest tertile group. Second, participants with markedly large values of hs-CRP were excluded. In addition, blood glucose and lipids were corrected for separately from Model C, as well as hs-CRP at baseline and aspirin use. After these exclusions and additional adjustments, the association remained significant in the highest tertile of hs-CRP variability but was attenuated and no longer significant in the middle tertile, suggesting that the effect was robust in the highest tertile but not in the middle tertile (Table V).

Moreover, VIM was used as a complementary measure of variability. The RCS plot showed a decrease in HR for cancer risk with increasing VIM of hs-CRP (Fig. 2B). Cox regression analysis was used to explore the association between VIM of hs-CRP and cancer risk. The results revealed that VIM demonstrated a similar association with cancer as CV. Furthermore, there was a significant association between higher VIM of hs-CRP and reduced cancer risk (Fig. S2).

Finally, the association between hs-CRP variability and the risk of mortality was assessed. In model A, high hs-CRP variability was significantly associated with an increased risk of death (HR, 0.81; 95% CI, 0.75-0.88; P<0.001). However, after adjustment for potential confounders in model C, the association was attenuated and no longer statistically significant (HR,0.92; 95% CI, 0.85-1.00; P=0.056), suggesting that the association observed in model A was largely explained by confounding factors (Table SIII). Nonetheless, competing risk plots suggest that high hs-CRP variability had a lower risk of cancer development and death than low variability (Fig. S3).

The findings of the present study indicate a significant association between high hs-CRP variability and a reduced cancer risk. Maintaining a low inflammatory state in the body is of notable importance for cancer prevention. Participants with high hs-CRP levels coupled with low variability exhibited a significantly increased risk of developing cancer, in the present study.

Previous studies have reported that hs-CRP serves not only as a prognostic factor for patients with cancer, but it is also associated with an increased risk of future cancer in seemingly healthy individuals (24). Suthahar et al (17) suggested that baseline hs-CRP was associated with a ~17% increase in cancer risk, and elevated hs-CRP over time was associated with an ~8% increase in cancer risk. Moreover, a study on the trajectory of hs-CRP changes suggested that maintaining a long-term, stable and low level of hs-CRP is positively meaningful for cancer prevention (20,25). The results of the present study are in line with this conclusion. Participants in a low inflammatory state (hs-CRP of <3 mg/l, or with a trajectory of low-stable) exhibited a significantly lower risk of developing cancer compared with that of participants in a high inflammatory state (hs-CRP of ≥3 mg/l). On the one hand, chronic hyperinflammatory states leads to cellular damage and DNA injury; on the other hand, the activation of pro-inflammatory pathways, such as the NF-κB pathway, may excessively inhibit surveillance by the immune system and clearance of malignant cells (26). These factors could potentially contribute to the eventual presence and growth of cancer.

Chronic inflammation is the result of incomplete resolution of acute inflammation and has long been considered one of the causes of cancer (27). Chronic inflammation can persist for several years or even throughout the lifetime of an individual, and in the long term, non-steroidal anti-inflammatory drugs are ineffective (28). One hypothesis is rooted in the concept of ‘unalamation’ proposed by Maddipati (29), which posits that inflammation in the absence of injury may arise from a reduction in anti-inflammatory mediators rather than an increase in pro-inflammatory mediators. In the present study, participants with consistently high hs-CRP levels exhibited the highest risk of developing cancer. Although the observational data lack direct immunological measurements to substantiate mechanistic claims, a notably speculative hypothesis is that a reduction in anti-inflammatory mediators may induce a state of immune dysregulation, manifesting as chronically stable, elevated hs-CRP. Consequently, the observed association should be considered exploratory, and residual confounding cannot be excluded. Future research incorporating specific biomarkers of immune function is essential to elucidate the underlying mechanisms and verify this speculation.

Previous studies exploring the association between hs-CRP and cancer have primarily focused on static hs-CRP levels, often overlooking its longitudinal variability. In response to inflammation or infection, the immune system activates the liver to release hs-CRP (30), resulting in a certain degree of physiological fluctuation, even in relatively healthy populations. Furthermore, adverse lifestyle factors such as smoking, obesity and chronic psychological stress are known to contribute to elevated hs-CRP levels (31,32). Therefore, predicting future cancer risk based solely on a single-point hs-CRP measurement is susceptible to confounding. The present study explored the association between hs-CRP variability and cancer incidence, and a notable observation was that higher hs-CRP variability was associated with a reduced risk of cancer. In an exploratory cross-classification analysis, participants exhibiting a ‘high’ hs-CRP level and low ‘variability’ profile had a ~42% higher observed risk of developing cancer compared with those with a 'low’ hs-CRP level and high ‘variability’ profile. This phenomenon corroborates previous epidemiological meta-analyses indicating a notable positive association between chronic inflammation, as characterized by elevated hs-CRP, and cancer incidence (33). In exploratory trajectory-stratified analyses, the inverse association between high hs-CRP variability and cancer risk was predominantly observed among individuals with a low-stable hs-CRP pattern. By contrast, no significant protective effect was detected in populations with decreasing or increasing hs-CRP trajectories; this may be due to sample size limitations, as the sample size for decreasing or increasing hs-CRP patterns were much lower than that of populations with low stable hs-CRP pattern.

Furthermore, the findings of the present study revealed a significant association between hs-CRP variability and the incidence of lung cancer. Compared with that of individuals with low hs-CRP variability, those with high hs-CRP variability exhibited a ~23% reduced risk of developing lung cancer. While elevated hs-CRP levels have been widely recognized as a relevant prognostic factor in patients with non-small cell lung cancer (32-34). In the context of immunotherapy for non-small cell lung cancer, hs-CRP has been identified as an independent prognostic marker, with a proposed mechanism involving its association with the immunosuppressive adenosine signature pathway (34). Furthermore, Feng et al (35) reported that hs-CRP inhibited the activation of lung macrophages induced by commensal bacteria in an FcγR2B-dependent manner, thereby impairing macrophage-mediated immune surveillance and facilitating the formation of the pre-metastatic niche in the lungs of tumor-bearing mice. Peripheral inflammation may influence the prognosis of patients with cancer by altering and shaping the tumor microenvironment (36). Hsu et al (37) reported that hs-CRP contributes to lung cancer development within the tumor microenvironment. However, research on the association between hs-CRP levels and the risk of lung cancer remains limited to date. The German Randomized Lung Cancer Screening Intervention trial reported that elevated hs-CRP levels were markedly associated with an increased risk of developing lung cancer (38). Moreover, the Kailuan Cohort Study also previously reported a notable association between high baseline hs-CRP levels and the occurrence of lung cancer (39). The findings of the present study further identified a significant association between hs-CRP variability and lung cancer. Additionally, a study assessing longitudinal changes in hs-CRP reported that participants with consistently low hs-CRP levels exhibited the lowest risk of developing cancer, whereas those with consistently high hs-CRP levels had the highest risk (17). Regarding potential mechanism underlying the association between hs-CRP and lung cancer incidence, systemic inflammation-reflected by elevated hs-CRP levels-may be associated with alterations in independent differentially methylated CpG loci, potentially contributing to lung carcinogenesis (40,41). However, given the observational nature of the present study and the absence of direct epigenetic or mechanistic measurements, these interpretations remain speculative. Further mechanistic research is warranted to substantiate the causal association between hs-CRP and lung cancer risk.

In subgroup analyses, the present study demonstrated that the inverse association between high hs-CRP variability and cancer incidence was more pronounced in older patients compared with that of middle-aged individuals, in smokers compared with that of non-smokers and in obese patients compared with that of non-obese participants. Both smoking and obesity are well-established inducers of a chronic systemic inflammatory state. As a marker of chronic inflammation, hs-CRP also reflects the biological aging status of the organism (42,43). A previous prospective cohort study reported that biological aging is frequently accompanied by low-grade inflammation, with aging of the immune system paralleled by elevated levels of pro-inflammatory markers such as hs-CRP (44). Previous reports have also proposed that biomarkers of immunosenescence may follow different age-related trajectories: Certain biomarkers increase progressively, while others remain relatively stable (45,46). hs-CRP is among the most commonly used markers of immune aging (47). In the present observational study, we hypothesize that stable hs-CRP levels may represent one manifestations of immunosenescence, whereas greater hs-CRP variability might reflect a more dynamic inflammatory state. Given that cancer is an age-related disease closely associated with aging, smoking and obesity (48), and that immunosenescence has been proposed as a key biological mechanism linking aging to cancer, we further hypothesized that hs-CRP variability, as a potential indicator of immune aging, may be associated with cancer risk. The findings of the present study demonstrated that higher hs-CRP variability was associated with lower cancer risk within these higher-risk groups; however, this inverse association should not be interpreted as causal or as evidence of a protective effect. As no direct measures of immunosenescence were available in the cohort, further exploration of whether hs-CRP variability is associated with immune aging was not possible. Future large-scale prospective studies could examine the relationships between hs-CRP variability and aging biomarkers or DNA methylation patterns and assess whether hs-CRP variability itself could serve as a marker of biological aging.

Furthermore, in the present observational study, an association between hs-CRP variability and cancer risk was identified. Although the present research does allow the elucidation of the biological mechanisms underlying this association, it may generate hypotheses for future research on inflammation dynamics and cancer prevention. The findings suggest that lower mean hs-CRP levels were associated with reduced cancer risk, which is consistent with prior evidence linking chronic inflammation to carcinogenesis.

A strength of the present study is the examination of longitudinal hs-CRP fluctuations in relation to cancer risk, an aspect that has received limited attention in previous research. To the best of our knowledge, few studies have directly evaluated the association between the magnitude of hs-CRP variability and incident cancer using repeated measurements over time. Treating variability as a fixed exposure is consistent with the study design and with previous research examining biomarker variability in relation to long-term health outcomes. However, the present study did not formally test whether CRP variability improves risk prediction beyond mean CRP or CRP trajectories. The cross-classification and trajectory-stratified analyses were exploratory and should be considered hypothesis-generating. While a time-dependent approach could theoretically capture changes in variability over time, it would require repeated recalculation of variability within moving time windows throughout the follow-up period. This approach is methodologically complex and may introduce instability, particularly when the number of hs-CRP measurements varies across individuals. Therefore, we consider the fixed-exposure approach is appropriate and aligns with the present research question. In addition, multiple potential confounders were adjusted for, including obesity and lifestyle factors, which strengthens the internal validity of the observed associations, although residual confounding cannot be fully excluded.

Nevertheless, the present study has several limitations. First, the sex distribution in the cohort is imbalanced, with a higher proportion of men than women. Although sex-stratified analyses were performed, residual confounding or selection bias associated with sex cannot be fully excluded. This concern is, however, partially mitigated by prior studies that did not observe effect modification by sex in the association between hs-CRP and cancer risk. Second, cancer development is multifactorial, and metabolic factors may partly confound the observed association. Although total cholesterol and triglycerides were adjusted for in sensitivity analyses, residual confounding by unmeasured or imperfectly measured metabolic factors remains possible. Third, inflammatory status may be influenced by conditions such as in autoimmune diseases. A cohort study on biological or targeted synthetic antirheumatic drugs and cancer reported that patients taking such drugs had an increased risk of developing cancer (49). However, information on autoimmune diseases was not available in the dataset used in the present study, precluding adjustment for this potential confounder. Fourth, as with any coefficient of variation, hs-CRP CV is mathematically dependent on the mean, which may introduce coupling between mean hs-CRP and its variability. To address this, trajectory-stratified analyses were performed and VIM was used, a variability metric less influenced by the mean. Nevertheless, a degree of residual dependence may persist, which is a recognized limitation of CV-based variability measures. Finally, the present study could not directly assess whether low hs-CRP variability reflects immunosenescence, as validated biomarkers or diagnostic criteria for immune aging were not measured in the cohort. As this was an observational cohort study, causality between hs-CRP variability and cancer risk cannot be established. Future studies incorporating biomarkers of immune aging and repeated inflammatory measures are warranted to clarify the relationships among hs-CRP variability, immunosenescence and cancer risk.

In conclusion, the present prospective cohort study demonstrated an inverse association between hs-CRP variability and the risk of cancer, particularly lung cancer. Participants with high hs-CRP levels and low variability demonstrated a higher cancer risk compared with those with low hs-CRP levels and high variability. However, due to the observational nature of the present study, causal inferences cannot be drawn. Further studies are needed to confirm these findings and to explore the underlying mechanisms.